China-linked group VAPOR PANDA used ARTEX AI agents running on DeepSeek to attack 200+ organizations in 11 days, breaching nine, CrowdStrike said
Cybersecurity company CrowdStrike said a Chinese state-linked hacking group it tracks as VAPOR PANDA used ARTEX, an open-source AI agent framework running mainly on the DeepSeek model, to attack more than 200 organizations in at least 14 countries over 11 days beginning on October 17. Targets included governments, pharmaceutical companies, semiconductor firms, defense contractors, think tanks and telecommunications providers. CrowdStrike confirmed that the attackers gained access and carried out further activity inside nine of those organizations, two of them U.S. pharmaceutical companies. At least 12 others faced credential attacks that CrowdStrike did not count as confirmed intrusions. According to CrowdStrike, the group also carried out extensive intrusions against the foreign ministries of two Latin American countries. CrowdStrike described this as the group's first observed use of AI-driven intrusion tools. DeepSeek has not publicly responded to the findings.
What the AI did
According to CrowdStrike, a single controller for the ARTEX agent framework directed break-in attempts against more than 200 organizations. ARTEX takes a goal from a human operator and then works through reconnaissance, exploitation and movement inside networks on its own. At each step it calls a language model to decide what to do next. The attackers ran it mainly on DeepSeek, plus a second paid commercial model that CrowdStrike could not identify. Recovered transcripts showed a human operator giving instructions in Chinese and stepping in at key decision points, while the agents did most of the work.
Added to the register November 8, 2026 · 2 sources
- Organizations with confirmed intrusions
- 9
- Deployer
- VAPOR PANDA
- Model
- DeepSeek, ARTEX
- Affected
- Governments and businesses worldwide, Foreign ministries of two Latin American countries, Two U.S. pharmaceutical companies
- When it happened
- October 17, 2026
- First reported
- Not stated in the sources
What this means for you
Could this affect you?
CrowdStrike said one AI-driven controller attacked more than 200 organizations in at least 14 countries in 11 days, including governments, drug makers, chip firms, defense contractors and telecom providers. Attackers using AI agents can reach many organizations like yours quickly and cheaply.
What to check
- Check whether your organization works in a sector CrowdStrike says was targeted, such as government, pharmaceuticals, semiconductors, defense, think tanks or telecommunications.
- Prioritise patching systems that can be reached from the internet.
- Watch for high-volume automated scanning, break-in attempts and credential attacks against your systems.
- Enforce strong authentication so that credential attacks are less likely to succeed.
Areas of your AI programme this touches
Timeline
- October 17, 2026Happened
- October 30, 2026CrowdStrike attributed the campaign to the China-based group VAPOR PANDA, which it said ran the open-source ARTEX agent framework primarily on DeepSeek, and confirmed that the attackers gained access and carried out follow-on activity inside nine targeted organizations in a campaign that began October 17 and reached at least 14 countries. Two of the compromised organizations were U.S. pharmaceutical companies, according to reporting on the findings, and the group also carried out extensive intrusions against the foreign ministries of two Latin American countries.[2]
Every fact and its source (7)
- Campaign startOctober 17“telecommunications providers beginning on October 17”[2]
- Attributed actorChinese state-linked hackers“Chinese state-linked hackers used artificial intelligence”[1]
- Attribution sourceCrowdStrike“according to the cybersecurity company CrowdStrike”[1]
- Campaign duration11 days“targeted more than 200 organizations in just 11 days”[1]
- Organizations targetedMore than 200“cyberattacks against more than 200 organizations around the world in just 11 days”[1]
- Scope of targetingMore than 200 organizations in at least 14 countries over 11 days“attack more than 200 organizations in at least 14 countries over 11 days this month”[2]
- Additional credential attacksAt least 12 other organizations“at least 12 others faced credential attacks that it did not count as confirmed intrusions”[2]
Sources
- China-linked hackers used AI to target 200 organizations, CrowdStrike saysnbcnews.com · October 30, 2026
- China-Linked Hackers Used DeepSeek AI Against 200 Organizationshngn.com · October 30, 2026
How this record is classified. Severity C (4/5): real harm at scale, or an autonomous system acting against people or institutions. OECD level: incident, an event in which an AI system led to actual harm. Evidence: Reported, meaning press reporting only.
The facts in At a glance and in Every fact and its source are quoted from the sources listed. The summary, What the AI did and What this means for you are written from those sources and checked against them automatically before publication. Records reflect what has been disclosed, not everything that has happened. Incidents are found through news feeds and the GDELT Project.
